|Publication number||US7478377 B2|
|Application number||US 10/918,996|
|Publication date||Jan 13, 2009|
|Filing date||Aug 16, 2004|
|Priority date||Jun 7, 2004|
|Also published as||US8196124, US20050283774, US20080307402|
|Publication number||10918996, 918996, US 7478377 B2, US 7478377B2, US-B2-7478377, US7478377 B2, US7478377B2|
|Inventors||Alexandre E. Eichenberger, Kai-Ting Amy Wang, Peng Wu|
|Original Assignee||International Business Machines Corporation|
|Export Citation||BiBTeX, EndNote, RefMan|
|Patent Citations (29), Non-Patent Citations (25), Referenced by (8), Classifications (6), Legal Events (5)|
|External Links: USPTO, USPTO Assignment, Espacenet|
This application is a continuation-in-part of commonly assigned, U.S. Non-Provisional patent application Ser. No. 10/862,483, entitled “SYSTEM AND METHOD FOR EFFICIENT DATA REORGANIZATION TO SATISFY DATA ALIGNMENT CONSTRAINTS,” filing date Jun. 7, 2004, now U.S. Pat. No. 7,386,842, which is incorporated herein by reference. This application is also related to U.S. Non-Provisional patent application Ser. No. 10/919,131, entitled “SYSTEM AND METHOD FOR SIMD CODE GENERATION FOR LOOPS WITH MIXED DATA LENGTHS,” U.S. Non-Provisional patent application Ser. No. 10/919,005, entitled “FRAMEWORK FOR GENERATING MIXED-MODE OPERATIONS IN LOOP-LEVEL SIMDIZATION,” U.S. Non-Provisional patent application Ser. No. 10/919,115, entitled “FRAMEWORK FOR INTEGRATED INTRA- AND INTER-LOOP AGGREGATION OF CONTIGUOUS MEMORY ACCESSES FOR SIMD VECTORIZATION,” and U.S. Non-Provisional patent application Ser. No. 10/918,879, entitled “FRAMEWORK FOR EFFICIENT CODE GENERATION USING LOOP PEELING FOR SIMD LOOP CODE WITH MULTIPLE MISALIGNED STATEMENTS,” which are all filed even date hereof, assigned to the same assignee, and incorporated herein by reference.
1. Technical Field
The present invention relates in general to a system and method for vectorizing loop code for execution on Single Instruction Multiple Datapath (SIMD) architectures that impose strict alignment constraints on the data.
2. Description of the Related Art
Multimedia extensions (MMEs) have become one of the most popular additions to general-purpose microprocessors. Existing multimedia extensions can be characterized as Single Instruction Multiple Datapath (SIMD) units that support packed fixed-length vectors. The traditional programming model for multimedia extensions has been explicit vector programming using either (in-line) assembly or intrinsic functions embedded in a high-level programming language. Explicit vector programming is time-consuming and error-prone. A promising alternative is to exploit vectorization technology to automatically generate SIMD codes from programs written in standard high-level languages.
Although vectorization has been studied extensively for traditional vector processors decades ago, vectorization for SIMD architectures has raised new issues due to several fundamental differences between the two architectures. See, e.g., Ren, Gang, et al. A Preliminary Study on the Vectorization of Multimedia Applications. In 16th International Workshop of Languages and Compilers for Parallel Computing. October 2003. To distinguish between the two types of vectorization, we refer to the latter as simdization. One such fundamental difference comes from the memory unit. The memory unit of a typical SIMD processor bears more resemblance to that of a wide scalar processor than to that of a traditional vector processor. In the VMX instruction set found on certain POWERPC microprocessors (produced by International Business Machines Corporation and Motorola, Inc.), for example, a load instruction loads 16-byte contiguous memory from 16-byte aligned memory, ignoring the last 4 bits of the memory address in the instruction. The same applies to store instructions. In this paper, architectures with alignment constraints refer to machines that support only loads and stores of register-length aligned memory.
There has been a recent spike of interest in compiler techniques to automatically extract SIMD parallelism from programs. See, e.g., LARSEN, Samuel, et al. Exploiting Superword Level Parallelism with Multimedia Instruction Sets. In Proceedings of SIGPLAN Conference on Programming Language Design and Implementation. June 2000, pp. 145-156; BIK, Aart, et al. Automatic Intra-Register Vectorization for the Intel Architecture. Int. J. of Parallel Programming. April 2002, vol. 30, no. 2, pp. 65-98; KRALL, Andreas, et al. Compilation Techniques for Multimedia Processors. Int. J. of Parallel Programming. August 2000, vol. 28, no. 4, pp. 347-361; SRERMAN, N., et al. A Vectorizing Compiler for Multimedia Extensions. Int. J. of Parallel Programming, August 2000, vol. 28, no. 4, pp. 363-400; LEE, Corinna G., et al. Simple Vector Microprocessors for Multimedia Applications. In Proceedings of International Symposium on Microarchitecture. 1998, pp. 25-36; and NAISHLOS, Dorit, et al. Vectorizing for a SIMD DSP Architecture. In Proceedings of International Conference on Compilers, Artchitectures, and Synthesis for Embedded Systems. October 2003, pp. 2-11. This upsurge was driven by the increasing prevalence of SIMD architectures in multimedia processors. Two principal techniques have been used, the traditional loop-based vectorization pioneered for vector supercomputers (e.g., ALLEN, John Randal, et al. Automatic Translation of Fortran Programs to Vector Form. ACM Transactions on Programming Languages and Systems. October 1987, vol. 4, pp. 491-542; and ZIMA, Hans, et al. Supercompilers for Parallel and Vector Computers. Reading, MA: Addison-Wesley/ACM Press, 1990. ISBN 0201175606.) and the unroll-and-pack approach first proposed by Larsen and Amarasinghe in LARSEN, Samuel, et al. Exploiting Superword Level Parallelism with Multimedia Instruction Sets. In Proceedings of SIGPLAN Conference on Programming Language Design and Implementation. June 2000, pp. 145-156.
The alignment constraints of SIMD memory units present a great challenge to automatic simdization. Consider the code fragment in
A very extensive discussion of alignment considerations is provided by LARSON, Samuel, et al. Increasing and Detecting Memory Address Congruence. In Proceedings of 11th International Conference on Parallel Architectures and Compilation Techniques. September 2002. However, LARSON is concerned with the detection of memory alignments and with techniques to increase the number of aligned references in a loop, whereas our work focuses on generating optimized SIMD codes in the presence of misaligned references. The two approaches are complementary. The use of loop peeling to align accesses was discussed in LARSON as well as in the aforementioned BIK reference. The loop peeling scheme is equivalent to the eager-shift policy with the restriction that all memory references in the loop must have the same misalignment. Even under this condition, our scheme has the advantage of generating simdized prologue and epilogue, which is the by-product of peeling from the simdized loop.
Direct code generation for misaligned references have been discussed by several prior works. The vectorization of misaligned loads and stores using the VIS instruction set is described in CHEONG, Gerald, et al. An Optimizer for Multimedia Instruction Sets. In Second SUIF Compiler Workshop. August 1997. The aforementioned BIK, et al. reference described a specific code sequence of aligned loads and shuffle to load memory references that cross cache line boundaries, which is implemented in Intel's compiler for SSE2. However, their method is not discussed in the context of general misalignment handling.
The VAST compiler, a commercial product by Crescent Bay Software, has some limited ability to simdize loops with multiple misaligned references, unknown loop bounds, and runtime alignments, and exploit the reuse when aligning a steam of contiguous memory. The VAST compiler, however, produces less than optimal simdized code, as its highly generalized scheme for handling mis-alignment can produce additional compilation overhead.
An interesting simdization scheme using indirect register accesses is discussed in the aforementioned NAISHLOS, et al. reference. However, their method is specific to the eLite processor, which supports more advanced vector operations (such as gather and scatter operations) than are available on typical MME processors. In SHIN, Jaewook, et al. Compiler-Controlled Caching in Superword Register Files for Multimedia Extension Architectures. In Proceedings of International Conference on Parallel Architectures and Compilation Techniques. September 2002, register packing and shifting instructions were used to exploit temporal and spatial reuse in vector registers. However, their work does not address alignment handling.
Another work that is of note, but which is in the area of compiling for distributed memory systems as opposed to SIMD architectures, is CRATTERJEE, Siddhartha, et al. Modeling Data-Parallel Programs with the Alignment-Distribution Graph. J. of Programming Languages. 1994, vol. 2, no. 3, pp. 227-258.
In the incorporated U.S. patent application Ser. No. 10/862,483 (hereinafter EICHENBERGER) a generic alignment handling framework that simdizes loops with arbitrary misalignments is disclosed. According to this framework, contiguous data accessed in a loop is viewed as streams, and aligning data to satisfy alignment constraints is modeled as shifting streams. Consider, for example, the (C-language) loop in
Although runtime alignment is handled in the framework of EICHENBERGER, it is not as efficient as the handling of compile time alignment. Due to code generation issues, stream shifts must be implemented as either stream shift left or stream shift right. In the presence of runtime alignment, the relative alignment of 2 streams is clearly unknown at compile time. In such cases, the approach taken in EICHENBERGER is to shift left each input memory stream to the leftmost position (register offset 0), perform the computation, and shift right the result to the store memory alignment. For example, in the loop in
In addition, existing technologies fail to adequately address the issue of data-length conversion in the generation of vectorized code for SIMD processors, where the source or destination data streams are misaligned with respect to each other. For example, one may write a loop that adds a vector of 16-bit “short” integers to a vector of 32-bit “long” integers to obtain a result that is a vector of 32-bit integer values (e.g., the case where b is an array of short integers and a and C are arrays of long integers in the loop of
Thus, what is needed is a method for automatically simdizing sequential program code into parallelized SIMD code in the presence of vector misalignments that are undefined at compile-time and where a conversion between datatypes of different lengths is needed. The present invention provides a solution to these and other problems, and offers other advantages over previous solutions.
The present invention generates loop code to execute on Single-Instruction Multiple-Datapath (SIMD) architectures, where the loop operates on datatypes having different lengths. The present invention efficiently realigns or shifts arbitrary streams to an arbitrary offset, regardless of whether the alignments or offsets are known at the compile time or not. This technique enables the application of advanced alignment optimizations to runtime alignment. Length conversion operations, for packing and unpacking data values, are included in the alignment handling framework. These operations are formally defined in terms of standard SIMD instructions that are readily available on various SIMD platforms. This allows sequential loop code operating on datatypes of disparate length to be transformed (“simdized”) into optimized SIMD code through a fully automated process.
The foregoing is a summary and thus contains, by necessity, simplifications, generalizations, and omissions of detail; consequently, those skilled in the art will appreciate that the summary is illustrative only and is not intended to be in any way limiting. Other aspects, inventive features, and advantages of the present invention, as defined solely by the claims, will become apparent in the non-limiting detailed description set forth below.
The present invention may be better understood, and its numerous objects, features, and advantages made apparent to those skilled in the art by referencing the accompanying drawings.
1. Preliminary Matters
In this section, an overview of the alignment handling framework of the incorporated EICHENBERGER application is provided, some of the key concepts that a preferred embodiment of the present invention is built upon are highlighted.
1.1. Alignment Constraints of a Valid Simdization
A valid vectorization for traditional vector units is only constrained by dependencies. This is no longer sufficient when simdizing for architectures with alignment constraints. Consider the loop a[i+2]=b[i+1]+c[i+3] in
The byte offset of data in a vector register is the key to expressing alignment constraints of a valid simdization. To capture the byte offset property within a loop, the abstraction of stream is introduced. There are two types of streams, a memory stream that represents the sequence of contiguous memory locations accessed by a memory reference over the lifetime of a loop, and a register stream that represents the sequence of consecutive registers produced by a SIMD operation over the lifetime of a loop.
Vector operations in a loop can be viewed as operations over streams. A vector load consumes a memory stream and produces a register stream. As data is read from memory in discrete chunks of 16 bytes, extra values may be introduced at the beginning and the end of a register stream. To indicate the boundaries of a register stream, the concept of a stream offset is introduced. The stream offset is defined as the byte offset of the first desired value in the first register of a register stream. Note that the stream offset of a register stream generated by of a vector load is dictated by the alignment of the input memory stream. The alignment of a memory stream is the alignment of the first element of the memory stream, namely the byte memory address of the first element modulo the byte width of the SIMD unit. For example, in
1.3 Stream Shift and Alignment Policies
A valid simdization requires the offsets of register streams involved in the computation to satisfy alignment constraints. In the presence of misalignments, this property is only achievable by judiciously inserting data reorganization operations to enforce the desired stream offset. The stream shift operation, vshiftstream(S,c), is introduced for this purpose. vshiftstream(S,c) shifts all values of a register stream S across consecutive registers of the stream to an offset of c. Note that, once the first element of a stream satisfies the alignment constraints, the rest of the stream also satisfies the alignment constraints. There are several policies to generate a valid simdization, each of which is discussed in greater detail in the incorporated EICHENBERGER application.
Zero-Shift Policy This policy shifts each misaligned load stream to 0, and shifts the register stream to be stored to the alignment of the store address. The simdization in
Eager-Shift Policy This policy shifts each load directly to the alignment of the store. The simdization in
Lazy-Shift Policy This policy is based on the eager-shift policy but pushes the shift as close to the root of the expression tree as possible. Consider the example of a[i+3]=b[i+1]+c[i+1]. This policy exploits the fact that b[i+1] and c[i+1] are relatively aligned, and thus can be safely operated on as is. Therefore, only the result of the add needs to be shifted so as to match the alignment of the store.
Dominant-Shift Policy This policy further reduces the number of stream shifts by shifting register streams to the most dominant stream offset in an expression tree.
2. Efficient Runtime Alignment Handling
In this section, an intuitive explanation of the approach applied in a preferred embodiment of the present invention is provided in Section 2.1. Then, the overall simdization framework in described in Section 2.2. Section 2.3 describes an algorithm to transform an arbitrary stream shift to a shift left, and the correctness of this algorithm is proved in Section 2.4.
Note that all examples in this section have compile-time alignments, for simplicity. It will be demonstrated, however, that the code generation algorithm employed in a preferred embodiment of the present invention never uses specific properties of their values. Thus, this algorithm is suitable for runtime alignment as well.
2.1. An Intuitive Example
Consider the initial example of a[i+2]=b[i+1]+c[i+3]. Given the data layout in
While this approach works well for stream offsets known at compile time, it does not work for runtime alignment for the following reason. Generally, shifting a stream from offset x to offset y corresponds to a right shift when x≦y and a left shift when x≧y. When generating codes for a right shift (as in
A key insight is to realize that this issue occurs because one is focusing on the first value of a stream. Consider, then, the case in which one focuses instead on a value before the first element, specifically on the value that is at offset zero of the register containing the first element after the shifting. For example, in
Suppose one focuses, now, on two new streams that are constructed by prepending a few values to the original b[i+1] and c[i+3] streams so that the new streams start at, respectively, b[−1] (value 316) and c (value 318). These new streams are shown in
By definition, the newly prepended streams will be shifted to offset zero. The shifted new streams yield the same sequence of registers as that produced by shifting the original stream (highlighted with dark grey box with light grey circle), as confirmed by visual inspection of the outcome of
2.2. Overview of the Simdization Framework
Now an overview of the overall simdization framework that the stream shift algorithm is based upon is provided. The framework that contains three phases:
Data reorganization phase. This phase takes, as input, an expression tree and generate a tree with generic SIMD operations as well as stream shifts that are inserted to satisfy the alignment constraints of the simdized operations. The augmented expression tree is called a data reorganization graph, such as the ones shown in
Shift normalization phase. This phase transforms all stream shifts in a data reorganization graph to stream left shifts. The output is still a data reorganization graph.
Code generation phase. This phase takes a data reorganization graph as input and maps generic SIMD operations and left shift operations to machine instructions. Specifically, the algorithm maps vshiftleft to native permutation instructions that “shifts” across a pair of vector registers, e.g., vec_perm on VMX. This phase also addresses issues such as resulting loads when shifting a stream, handling unknown loop counts and multiple statements, and efficient prologue/epilogue generation.
Of primary importance to a preferred embodiment of the present invention is the shift normalization phase that converts arbitrary stream shifts to stream shift left. The other two phases function essentially as described in the incorporated EICHENBERGER reference. Therefore, the rest of the section will focus on the shift normalization algorithm.
2.3. Transforming a Stream Shift to Shift Left
As illustrated in Section 2.1, an arbitrary stream shift, vshiftstream(S,x), can be converted to a stream shift left by shifting a derived stream, one that starts exactly x bytes before the first value of S, to offset zero. For the ease of description, introduce two operators are introduced: Prepend(S,x), which prepends x bytes at the beginning of stream S, and Skip(S, x), which skips the first x bytes of stream S. Using these two new operators, the above statement can be restated as,
Consider now the Skip operation in Equation (1). In this equation, one always skips y<V bytes from a register stream with offset zero. (By definition, vshiftstream(Prepend(S,y),0) has an offset of zero.) Since a register stream is a sequence of V byte wide registers, skipping strictly less than V bytes from offset 0 never skips a whole register and thus has no effect on which registers to produce in a register stream. It only serves for bookkeeping purposes to maintain that the transformed stream has the same offset and length properties as the original stream. Thus, it can be safely ignored during the code generation phase.
For the Prepend operation in Equation (1), there are two cases to consider. When prepending to a memory stream, one simply subtracts x bytes from the memory address. Directly prepending to a register stream is not feasible. However, one can propagate Prepend operations to the leaves of an expression tree until memory streams are reached. Below we describe how to perform Prepend for each type of nodes in a data reorganization graph.
2.4. Proof of Correctness
In this section, the correctness of two important equations in Section 2.3 is proved. The first theorem demonstrates the validity of sinking a prepend past a stream shift, i.e., Equation (4).
Theorem 1. The following equality holds for an arbitrary stream S and V-byte wide SIMD unit:
Proof. As a preliminary step, the properties of Prepend(S,x) are first defined as functions of those of S. As x bytes are added at the head of the stream, the offset and length of the prepended stream becomes (Offset(S)−x)modV and Length(S)+x, respectively. Similarly, the offset and length properties of VSHIFTSTREAM(S,to) becomes to and Length(S), respectively.
The equality of the two streams is now proven by demonstrating that they derive from the same register stream and have the same offset and length properties.
Offset(Prepend(VSHIFTSTREAM(S, = (Offset(VSHIFTSTREAM(S, to), x)) to)) − x) mod V = (to − x) mod V Length(Prepend(VSHIFTSTREAM(S, = Length(VSHIFTSTREAM(S, to), x)) to)) + x = Length(S) + x Offset(VSHIFTSTREAM(Prepend(S, = (to − x) mod V x), (to − x) mod V)) Length(VSHIFTSTREAM(prepend(S, = Length(S) + x x), (to − x) mod V) □
Using Theorem 1, the validity of transforming arbitrary stream shift to stream shift left, i.e., Equation (1), can be demonstrated.
Theorem 2. A register stream S with arbitrary offset can be shifted to an arbitrary target offset to by (1) prepending to bytes to S, (2) shifting the resulting stream to offset zero, and (3) skipping the first to values of the resulting stream.
Proof. Since Skip is a reciprocal of Prepend, the following S1=Skip(Prepend(S1,x),x) holds for any stream S1 Let S1 be VSHIFTSTREAM(S,to), it can be stated that
Using Theorem 1 to permute Prepend and VSHIFTSTREAM, one obtains,
3. Alignment Handling for Length Conversion
For the purpose of describing a preferred embodiment of the present invention, the term “length conversion operation” is defined as any operation that converts a stream into a stream of a different length. The most common length conversion operations are type conversions between data of different sizes. When processing typecasts (as in the C programming language, which supports long and short “int” types, for example), the primary focus is on the data reorganization aspect of type conversion. Further data processing beyond data reorganization, such as sign extensions or floating point conversions, for example, can be handled by additional non-length-conversion vector operations.
Two types of length conversion operations are defined.
Length conversions have a significant impact on the simdization framework. The rest of the section explains how to incorporate them into the three phases of the framework.
3.1. Data Reorganization Phase
The scaling effect of length conversion on stream offsets affects all but the least optimized Zero-shift alignment policy. The Zero-shift policy is not affected by length conversions because the scaling impact on an offset of value 0 still results in an offset of 0. As defined in Section 1.3 and the incorporated EICHENBERGER application (in the absence of length conversions), the Eager-shift policy shifts each misaligned load stream directly to the alignment of the store. In the presence of length conversions, the store alignment is still propagated to load nodes, but its alignment value is scaled up/down by the packing/unpacking factors when traversing length conversion nodes. (Since stream offsets are propagated from the store to the loads, in the reverse direction of the data flow, Equations (5) and (6) are used in the reverse direction.) Similarly in the Lazy-shift policy, the target shift offset is also scaled when moving stream shifts past a pack/unpack node, i.e.,
3.2 Stream Shift Normalization Phase
In the normalization phase, the implementation of Prepend is augmented to accommodate length conversions. Consider the example of a[i+5]=(short)b[i+3] where a is an array of short and b is an array of integer.
The solution employed by a preferred embodiment of the present invention is to eliminate the dependence on address truncation that occurs at the memory operations by explicitly performing the desired truncation at the input stream of the stream shift operation. The solution is to further prepend the already prepended input stream of vshiftstream to the value that is at offset zero in the input register. For example, since the prepended input stream is b[−2] (values 502, 514), the stream is further prepended so that b[−8] (values 503, 516) is now the initial stream value. As can be visually seen from the figure by tracing the b[−8] value (value 516), the memory load truncation that used to occur at the load operation now occurs at the vicinity of shift stream operation 518, and is of the correct amount. The result may then be directly stored in the proper memory location for array “a” (store operation 520), with the result being truncated such that value 521 becomes the first value stored in array “a” (at value 522).
Theorem 3. In the presence of length conversion, a stream S can be shifted to an arbitrary target offset to by (1) prepending stream S by to+(Offset(S)−to)modV bytes, (2) skipping the first (Offset(S)−to)modV values of the resulting stream, (3) shifting the resulting stream to offset zero, and (4) skip to bytes of the shifted stream.
Note that, in Theorem 3, steps (1) and (3) produce streams of offset zero and in steps (2) and (4) the skip bytes are less than V. Therefore, the skips are only for bookkeeping purposes and can be safely ignored by code generation. Theorem 3 may be proved mathematically in a similar fashion to Theorem 2.
3.3. Code Generation Phase
During the code generation, vpack and vunpack are mapped to native permutation instructions. Some multimedia platforms even provide native packing and unpacking instructions to perform these operations, e.g., VMX. For machines where length conversions and stream shifts are both mapped to some sort of permutation instruction, the total number of extra permutation operations can be further reduced by merging together pairs of consecutive length conversion and stream shift.
4. Method Overview and Example Hardware Platform Description
PCI bus 714 provides an interface for a variety of devices that are shared by host processor(s) 700 and Service Processor 716 including, for example, flash memory 718. PCI-to-ISA bridge 735 provides bus control to handle transfers between PCI bus 714 and ISA bus 740, universal serial bus (USB) functionality 745, power management functionality 755, and can include other functional elements not shown, such as a real-time clock (RTC), DMA control, interrupt support, and system management bus support. Nonvolatile RAM 720 is attached to ISA Bus 740. Service Processor 716 includes JTAG and I2C buses 722 for communication with processor(s) 700 during initialization steps. JTAG/I2C buses 722 are also coupled to L2 cache 704, Host-to-PCI bridge 706, and main memory 708 providing a communications path between the processor, the Service Processor, the L2 cache, the Host-to-PCI bridge, and the main memory. Service Processor 716 also has access to system power resources for powering down information handling device 701.
Peripheral devices and input/output (I/O) devices can be attached to various interfaces (e.g., parallel interface 762, serial interface 764, keyboard interface 768, and mouse interface 770 coupled to ISA bus 740. Alternatively, many I/O devices can be accommodated by a super I/O controller (not shown) attached to ISA bus 740.
In order to attach computer system 701 to another computer system to copy files over a network, LAN card 730 is coupled to PCI bus 710. Similarly, to connect computer system 701 to an ISP to connect to the Internet using a telephone line connection, modem 775 is connected to serial port 764 and PCI-to-ISA Bridge 735.
While the computer system described in
One of the preferred implementations of the invention is a client application, namely, a set of instructions (program code) or other functional descriptive material in a code module that may, for example, be resident in the random access memory of the computer. Until required by the computer, the set of instructions may be stored in another computer memory, for example, in a hard disk drive, or in a removable memory such as an optical disk (for eventual use in a CD ROM) or floppy disk (for eventual use in a floppy disk drive), or downloaded via the Internet or other computer network. Thus, the present invention may be implemented as a computer program product for use in a computer. In addition, although the various methods described are conveniently implemented in a general purpose computer selectively activated or reconfigured by software, one of ordinary skill in the art would also recognize that such methods may be carried out in hardware, in firmware, or in more specialized apparatus constructed to perform the required method steps. Functional descriptive material is information that imparts functionality to a machine. Functional descriptive material includes, but is not limited to, computer programs, instructions, rules, facts, definitions of computable functions, objects, and data structures.
While particular embodiments of the present invention have been shown and described, it will be obvious to those skilled in the art that, based upon the teachings herein, that changes and modifications may be made without departing from this invention and its broader aspects. Therefore, the appended claims are to encompass within their scope all such changes and modifications as are within the true spirit and scope of this invention. Furthermore, it is to be understood that the invention is solely defined by the appended claims. It will be understood by those with skill in the art that if a specific number of an introduced claim element is intended, such intent will be explicitly recited in the claim, and in the absence of such recitation no such limitation is present. For non-limiting example, as an aid to understanding, the following appended claims contain usage of the introductory phrases “at least one” and “one or more” to introduce claim elements. However, the use of such phrases should not be construed to imply that the introduction of a claim element by the indefinite articles “a” or “an” limits any particular claim containing such introduced claim element to inventions containing only one such element, even when the same claim includes the introductory phrases “one or more” or “at least one” and indefinite articles such as “a” or “an”; the same holds true for the use in the claims of definite articles.
|Cited Patent||Filing date||Publication date||Applicant||Title|
|US4710872||Aug 7, 1985||Dec 1, 1987||International Business Machines Corporation||Method for vectorizing and executing on an SIMD machine outer loops in the presence of recurrent inner loops|
|US5247645||Mar 12, 1991||Sep 21, 1993||International Business Machines Corporation||Dynamic memory mapper which supports interleaving across 2N +1, 2.sup.NN -1 number of banks for reducing contention during nonunit stride accesses|
|US5349665||Jan 13, 1993||Sep 20, 1994||Nec Corporation||Compiler vectorizing system|
|US5440687||Jan 29, 1993||Aug 8, 1995||International Business Machines Corporation||Communication protocol for handling arbitrarily varying data strides in a distributed processing environment|
|US5481746||Mar 29, 1994||Jan 2, 1996||Cray Research, Inc.||Vector shift functional unit for successively shifting operands stored in a vector register by corresponding shift counts stored in another vector register|
|US5522074||Dec 14, 1993||May 28, 1996||Nec Corporation||Vectorization system for vectorizing loop containing condition induction variables|
|US5802375||Nov 23, 1994||Sep 1, 1998||Cray Research, Inc.||Outer loop vectorization|
|US5933650||Oct 9, 1997||Aug 3, 1999||Mips Technologies, Inc.||Alignment and ordering of vector elements for single instruction multiple data processing|
|US5996057||Apr 17, 1998||Nov 30, 1999||Apple||Data processing system and method of permutation with replication within a vector register file|
|US6113650||Jul 1, 1998||Sep 5, 2000||Nec Corporation||Compiler for optimization in generating instruction sequence and compiling method|
|US6202141||Jun 16, 1998||Mar 13, 2001||International Business Machines Corporation||Method and apparatus for performing vector operation using separate multiplication on odd and even data elements of source vectors|
|US6266758||Mar 5, 1999||Jul 24, 2001||Mips Technologies, Inc.||Alignment and ordering of vector elements for single instruction multiple data processing|
|US6334176||Apr 17, 1998||Dec 25, 2001||Motorola, Inc.||Method and apparatus for generating an alignment control vector|
|US6484255||Sep 20, 1999||Nov 19, 2002||Intel Corporation||Selective writing of data elements from packed data based upon a mask using predication|
|US6550059||Oct 4, 1999||Apr 15, 2003||Advanced Micro Devices, Inc.||Method for generating optimized vector instructions from high level programming languages|
|US7020873||Jun 21, 2002||Mar 28, 2006||Intel Corporation||Apparatus and method for vectorization of detected saturation and clipping operations in serial code loops of a source program|
|US7168069 *||Jul 12, 2000||Jan 23, 2007||Stmicroelectronics, Inc.||Dynamic generation of multimedia code for image processing|
|US7219212||Feb 25, 2005||May 15, 2007||Tensilica, Inc.||Load/store operation of memory misaligned vector data using alignment register storing realigned data portion for combining with remaining portion|
|US7275147 *||Mar 31, 2003||Sep 25, 2007||Hitachi, Ltd.||Method and apparatus for data alignment and parsing in SIMD computer architecture|
|US20030120889 *||Dec 21, 2001||Jun 26, 2003||Patrice Roussel||Unaligned memory operands|
|US20030167460||Feb 26, 2002||Sep 4, 2003||Desai Vipul Anil||Processor instruction set simulation power estimation method|
|US20030204840||Apr 30, 2002||Oct 30, 2003||Youfeng Wu||Apparatus and method for one-pass profiling to concurrently generate a frequency profile and a stride profile to enable data prefetching in irregular programs|
|US20040001066||Jun 21, 2002||Jan 1, 2004||Bik Aart J.C.||Apparatus and method for vectorization of detected saturation and clipping operations in serial code loops of a source program|
|US20040003381||Jun 19, 2003||Jan 1, 2004||Fujitsu Limited||Compiler program and compilation processing method|
|US20040006667||Jun 21, 2002||Jan 8, 2004||Bik Aart J.C.||Apparatus and method for implementing adjacent, non-unit stride memory access patterns utilizing SIMD instructions|
|US20040025150||Jul 31, 2003||Feb 5, 2004||Taketo Heishi||Compiler, compiler apparatus and compilation method|
|US20040098709||Nov 12, 2003||May 20, 2004||Nec Corporation||Method, apparatus, and computer program for generating SIMD instruction sequence|
|US20050097301||Oct 29, 2003||May 5, 2005||International Business Machines Corporation||Vectorization in a SIMdD DSP architecture|
|US20050289529||Jun 24, 2004||Dec 29, 2005||Yoav Almog||Method and apparatus to vectorize multiple input instructions|
|1||Allen et al., "Automatic Translation of Fortran Programs to Vector Form," ACM Transactions on Programming Languages and Systems, v. 9, No. 4, Oct. 1987, p. 491-542.|
|2||AltiVec(TM) Technology Programming Interface Manual, 1999, 262 pages.|
|3||Arafeh et al., "Vectorization and Parallelization Interactive Assistant," 1988, ACM 0-89791-260-8/88/0002/0573, pp. 573-577.|
|4||Bik et al., "Automatic Intra-Register Vectorizatoin for the Intel Architecture," International Journal of Parallel Programming, vol. 30, No. 2, Apr. 2002 (p. 65-98).|
|5||Chatterjee et al., "Modeling Data-Parallel Programs with the Alignment-Distribution Graph," 6th Annual Workshop on Languages and Compilers for Parallelism, Aug. 1993, 28 pages.|
|6||Cheong et al., "An Optimizer for Multimedia Instruction Sets," Stanford University, Aug. 14, 1997.|
|7||Cukic et al., "Automatic Array Alignment as a Step in Hierarchical Program Transformation," Apr. 1994, Proceedings of the Eighth International Parallel Processing Symposium.|
|8||Eichenberger et al., Vectorization for SIMD Architectures with Alignment Constraints, PLD1'04, Jun. 9-11, 2004, p. 82-93.|
|9||Krall et al., "Compilation Techniques for Multimedia Processors," 14 pages.|
|10||Larsen et al., "Exploiting Superword Level Parallelism with Multimedia Instruction Sets," 2000, PLDI 2000, pp. 145-156.|
|11||Larsen et al., "Increasing and Detecting Memory Address Congruence," Proceedings of 11th International Conference on Parallel Architectures and Compilation Techniques, Sep. 2002, 11 pages.|
|12||Lee et al., "Automatic Data and Computation Decomposition on Distributed Memory Parallel Computers," Jan. 2002, ACM Transactions on Programming Languages and Systems, vol. 24, No. 1, pp. 1-50.|
|13||Lee et al., "Simple Vector Microprocessors for Multimedia Applications," Preceedings of the 31st Annual International Symposium on Microarchitecture, Dec. 1998 (12 pages).|
|14||Levine et al., "A Comparative Study of Automatic Vectorizing Compilers," 1991, Parallel Computing, vol. 17.|
|15||Lorenz et al., "Energy Aware Compilation for DSPs with SIMD Instructions," 2002, LCTES'02-SCOPES'02, pp. 94-101.|
|16||McKenney et al., "Generating Parallel Code for SIMD Machines," ACM Letters on Programming Languages and Systems, vol. 1, No. 1, Jan. 1992, pp. 59-73.|
|17||Naishlos et al., "Vectorizing for a SIMdD DSP Architecture," Cases '03, Oct. 30-Nov. 1, 2003, ACM 1-58113-676-5/03/0010, pp. 2-11.|
|18||Naishlos, "Vectorizing for a SIMdD DSP Architecture," Cases '03, Oct. 30-Nov. 1, 2003, ACM, pp. 2-11.|
|19||O'Brien, "Predictive Commoning: A method of optimizing loops containing references to consecutive array elements," 12 pages.|
|20||Office action for U.S. Appl. No. 10/919,005, made final, mailed Aug. 6, 2008.|
|21||Pajuelo et al., "Speculative Dynamic Vectorization," 2002, IEEE 1063-6897/02, pp. 271-280.|
|22||Ren et al., "A Preliminary Study on the Vectorization of Multimedia Applications for Multimedia Extensions," 15 pages.|
|23||Scarborough et al., "A Vectorizing Fortran Compiler," Mar. 1986, IBM Journal of Research and Development, vol. 30, Issue 2, pp. 163-171.|
|24||Shin et al., "Compiler-Controlled Caching in Superword Register Files for Multimedia Extension Architectures," 11 pages.|
|25||Sreraman et al., "A Vectorizing Compiler for Multimedia Extensions," International Journal of Parallel Programming, 2000 (40 pages).|
|Citing Patent||Filing date||Publication date||Applicant||Title|
|US8136105 *||Sep 29, 2006||Mar 13, 2012||International Business Machines Corporation||Method to exploit superword-level parallelism using semi-isomorphic packing|
|US9244677 *||Sep 28, 2012||Jan 26, 2016||Intel Corporation||Loop vectorization methods and apparatus|
|US9268541||Mar 15, 2013||Feb 23, 2016||Intel Corporation||Methods and systems to vectorize scalar computer program loops having loop-carried dependences|
|US9348592||Dec 22, 2011||May 24, 2016||Intel Corporation||Apparatus and method for sliding window data access|
|US20060195828 *||Feb 27, 2006||Aug 31, 2006||Kabushiki Kaisha Toshiba||Instruction generator, method for generating instructions and computer program product that executes an application for an instruction generator|
|US20080127144 *||Sep 29, 2006||May 29, 2008||International Business Machines Corporation||Method to exploit superword-level parallelism using semi-isomorphic packing|
|US20140096119 *||Sep 28, 2012||Apr 3, 2014||Nalini Vasudevan||Loop vectorization methods and apparatus|
|WO2013095581A1 *||Dec 22, 2011||Jun 27, 2013||Intel Corporation||Apparatus and method for sliding window data access|
|U.S. Classification||717/150, 717/160|
|Cooperative Classification||G06F8/4452, G06F8/447|
|Dec 20, 2004||AS||Assignment|
Owner name: INTERNATIONAL BUSINESS MACHINES CORPORATION, NEW Y
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:EICHENBERGER, ALEXANDRE E.;WANG, KAI-TING A.;WU, PENG;REEL/FRAME:015477/0221;SIGNING DATES FROM 20040811 TO 20040812
|Jun 29, 2012||FPAY||Fee payment|
Year of fee payment: 4
|Aug 26, 2016||REMI||Maintenance fee reminder mailed|
|Jan 13, 2017||LAPS||Lapse for failure to pay maintenance fees|
|Mar 7, 2017||FP||Expired due to failure to pay maintenance fee|
Effective date: 20170113